A generic bandwidth allocation scheme for multimedia substreams in adaptive networks using genetic algorithms

Abstract

In this paper we propose a generic call admission control scheme based on genetic algorithms. The scheme is applied to a wireless network using an adaptive resource allocation framework. A multimedia connection is represented in terms of three substreams (video, audio and data) each with a pre-specified range of acceptable quality of service (QoS) requirements. This range of requirements make the application adaptive in the sense that each of its substreams specify to the network a number of acceptable QoS levels instead of just a single one. For example, MPEG-based video can be transmitted at different levels of varying qualities. As resource availability in the wireless network changes, the control scheme selects the best possible QoS level that each substream can obtain for a certain period of time. In case of congestion, the control scheme attempts to free up some wireless resources by degrading the QoS of existing calls to a lesser level in order to accommodate new calls. This is done, however, under the constraint of achieving maximum utilization of the resources; and sharing them fairly among the calls. The degradation in the QoS is limited to a minimum value for each multimedia substream defined in a user-defined profile. From the user perspective, the perception of the QoS degradation is very graceful and happens only during over-load periods. The network services, on the other hand, are greatly enhanced due to the fact that the call blocking probability is significantly decreased. Genetic algorithms have been used to solve the optimization problem. Simulation results for a wireless ATM network demonstrate that the proposed scheme performs well in terms of increasing the number of calls admitted while effectively utilizing the available bandwidth fairly among the calls.